devrites 5.12.0 → 5.13.0

package/engine/internal/overhaul/score/run_test.go

@@ -0,0 +1,405 @@
+package score
+
+import (
+	"bytes"
+	"encoding/json"
+	"fmt"
+	"maps"
+	"math/big"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"github.com/devrites/devrites/internal/overhaul/ovio"
+)
+
+var doms = []string{"correctness", "security", "reliability", "performance", "tests", "architecture", "ux", "operations"}
+
+func ctl(i int, d string, w any, lanes ...string) map[string]any {
+	if len(lanes) == 0 {
+		lanes = []string{"backend"}
+	}
+	return map[string]any{"id": fmt.Sprintf("C-%03d", i), "domain": d, "lanes": lanes, "weight": w, "hard_gate": false}
+}
+
+func rubricOf(controls []map[string]any) map[string]any {
+	return map[string]any{"schema": "overhaul.rubric/1", "rev": 1, "lanes": []string{"frontend", "backend"},
+		"controls": controls, "domains": map[string]any{"correctness": 20, "security": 20, "reliability": 15,
+			"performance": 15, "tests": 10, "architecture": 10, "ux": 5, "operations": 5}}
+}
+
+func resOf(status map[string]string) map[string]any {
+	r := map[string]any{}
+	for k, v := range status {
+		r[k] = map[string]any{"status": v, "evidence": []string{"EV-1"}}
+	}
+	return map[string]any{"schema": "overhaul.results/1", "subject": "candidate", "results": r}
+}
+
+func passing() map[string]any {
+	g := map[string]any{}
+	for _, id := range auditGates {
+		g[id] = map[string]any{"status": "PASS", "evidence": []string{"EV-g"}}
+	}
+	return map[string]any{"mode": "audit", "gates": g}
+}
+
+func allPass(cs []map[string]any) map[string]string {
+	st := map[string]string{}
+	for _, c := range cs {
+		st[c["id"].(string)] = "PASS"
+	}
+	return st
+}
+
+// norm round-trips v through JSON with UseNumber, exactly as file input arrives.
+func norm(t *testing.T, v any) map[string]any {
+	t.Helper()
+	b, err := json.Marshal(v)
+	if err != nil {
+		t.Fatal(err)
+	}
+	dec := json.NewDecoder(bytes.NewReader(b))
+	dec.UseNumber()
+	var m map[string]any
+	if err := dec.Decode(&m); err != nil {
+		t.Fatal(err)
+	}
+	return m
+}
+
+func tryScore(t *testing.T, r, s, g any) (map[string]any, error) {
+	t.Helper()
+	out, err := score(norm(t, r), norm(t, s), norm(t, g))
+	if err != nil {
+		return nil, err
+	}
+	return norm(t, out), nil
+}
+
+func mustScore(t *testing.T, r, s, g any) map[string]any {
+	t.Helper()
+	out, err := tryScore(t, r, s, g)
+	if err != nil {
+		t.Fatalf("score: %v", err)
+	}
+	return out
+}
+
+func get(v any, path string) string { return fmt.Sprint(ovio.Get(v, path)) }
+
+func exact(t *testing.T, v any, path string) *big.Rat {
+	t.Helper()
+	r, ok := new(big.Rat).SetString(get(v, path))
+	if !ok {
+		t.Fatalf("%s is not exact: %v", path, ovio.Get(v, path))
+	}
+	return r
+}
+
+func base() []map[string]any {
+	var b []map[string]any
+	for i, d := range doms {
+		b = append(b, ctl(i, d, 5, "frontend", "backend"))
+	}
+	return b
+}
+
+func workedExample() ([]map[string]any, map[string]string) {
+	ex := append(base(), ctl(10, "correctness", 5), ctl(11, "correctness", 5), ctl(12, "correctness", 3), ctl(13, "correctness", 1))
+	st := allPass(ex)
+	st["C-013"] = "FAIL"
+	return ex, st
+}
+
+func laneMasking() []map[string]any {
+	l := base()
+	for k := range 10 {
+		l = append(l, ctl(30+k, "correctness", 5, "backend"))
+	}
+	return append(l, ctl(40, "correctness", 5, "frontend"))
+}
+
+// 1. Worked example: correctness 5,5,5,3 PASS + 1 FAIL -> D=180/19, Q=18800/1900.
+func TestWorkedExample(t *testing.T) {
+	ex, st := workedExample()
+	out := mustScore(t, rubricOf(ex), resOf(st), passing())
+	if got := get(out, "global.domains.correctness.exact"); got != "180/19" {
+		t.Fatalf("correctness = %s", got)
+	}
+	if got := get(out, "global.Q.exact"); got != "188/19" {
+		t.Fatalf("Q = %s", got)
+	}
+	if got := get(out, "global.Q.decimal"); got != "9.894736842105" {
+		t.Fatalf("decimal = %s", got)
+	}
+	if v := get(out, "readiness_verdict"); v != "MEETS_TARGETS" {
+		t.Fatalf("verdict = %s", v)
+	}
+}
+
+// 2. Same numbers, failing control is a hard gate -> NOT_READY despite Q 9.89.
+func TestHardGateBlocks(t *testing.T) {
+	ex, st := workedExample()
+	ex[len(ex)-1]["hard_gate"] = true
+	out := mustScore(t, rubricOf(ex), resOf(st), passing())
+	if get(out, "readiness_verdict") != "NOT_READY" || get(out, "gates.G-MANDATORY") != "FAIL" {
+		t.Fatalf("verdict %s, G-MANDATORY %s", get(out, "readiness_verdict"), get(out, "gates.G-MANDATORY"))
+	}
+	if got := get(out, "global.Q.exact"); got != "188/19" {
+		t.Fatalf("Q = %s", got)
+	}
+}
+
+// 3. Q exactly 9.696 fails 9.7 at full precision.
+func TestFullPrecisionThreshold(t *testing.T) {
+	var cs []map[string]any
+	st := map[string]string{}
+	i := 100
+	for _, d := range doms {
+		n := 1
+		if d == "correctness" || d == "security" {
+			n = 250
+		}
+		for k := range n {
+			cs = append(cs, ctl(i, d, 1, "frontend", "backend"))
+			st[fmt.Sprintf("C-%03d", i)] = "PASS"
+			if n == 250 && k < 19 {
+				st[fmt.Sprintf("C-%03d", i)] = "FAIL"
+			}
+			i++
+		}
+	}
+	out := mustScore(t, rubricOf(cs), resOf(st), passing())
+	if got := get(out, "global.Q.exact"); got != "1212/125" {
+		t.Fatalf("Q = %s", got)
+	}
+	row := ovio.List(out["thresholds"])[0]
+	if get(row, "scope") != "global" || get(row, "result") != "FAIL" || get(out, "readiness_verdict") != "NOT_READY" {
+		t.Fatalf("row %v verdict %s", row, get(out, "readiness_verdict"))
+	}
+}
+
+// 4. Zero denominator: an applicable domain whose only control is N/A -> NOT_ASSESSABLE, never 10.
+func TestZeroDenominatorNotAssessable(t *testing.T) {
+	b := base()
+	z := make([]map[string]any, len(b))
+	for i, c := range b {
+		z[i] = maps.Clone(c)
+	}
+	z[6]["na"] = map[string]any{"reason": "no UI", "evidence": []string{"EV-2"}, "approved": true}
+	out := mustScore(t, rubricOf(z), resOf(allPass(b)), passing())
+	if ovio.Get(out, "global.Q") != nil || get(out, "readiness_verdict") != "NOT_ASSESSABLE" {
+		t.Fatalf("Q %v verdict %s", ovio.Get(out, "global.Q"), get(out, "readiness_verdict"))
+	}
+}
+
+// 5. Global 59/6 and backend 10 cannot mask frontend 9.0; shared controls count once globally.
+func TestLaneMasking(t *testing.T) {
+	l := laneMasking()
+	st := allPass(l)
+	st["C-040"] = "FAIL"
+	out := mustScore(t, rubricOf(l), resOf(st), passing())
+	if row := ovio.List(out["thresholds"])[0]; get(row, "result") != "PASS" || get(out, "global.Q.exact") != "59/6" {
+		t.Fatalf("global row %v Q %s", row, get(out, "global.Q.exact"))
+	}
+	if get(out, "lanes.frontend.Q.exact") != "9/1" || get(out, "lanes.backend.Q.exact") != "10/1" {
+		t.Fatalf("lanes %v", out["lanes"])
+	}
+	laneRow := "missing"
+	for _, row := range ovio.List(out["thresholds"]) {
+		if get(row, "scope") == "lane frontend" {
+			laneRow = get(row, "result")
+		}
+	}
+	if laneRow != "FAIL" {
+		t.Fatalf("lane frontend threshold row = %s", laneRow)
+	}
+	if get(out, "readiness_verdict") != "NOT_READY" {
+		t.Fatalf("verdict %s", get(out, "readiness_verdict"))
+	}
+	if get(out, "counts.controls") != "19" || get(out, "counts.lane_appearances") != "27" {
+		t.Fatalf("counts %v", out["counts"])
+	}
+}
+
+// 6. PASS without evidence counts as UNKNOWN; UNKNOWN earns zero, not N/A.
+func TestPassWithoutEvidenceIsUnknown(t *testing.T) {
+	b := base()
+	r := resOf(allPass(b))
+	r["results"].(map[string]any)["C-000"].(map[string]any)["evidence"] = []string{}
+	out := mustScore(t, rubricOf(b), r, passing())
+	if get(out, "status_by_control.C-000") != "UNKNOWN" || get(out, "global.domains.correctness.exact") != "0/1" {
+		t.Fatalf("status %s correctness %s", get(out, "status_by_control.C-000"), get(out, "global.domains.correctness.exact"))
+	}
+	if len(ovio.List(out["downgrades"])) != 1 {
+		t.Fatalf("downgrades %v", out["downgrades"])
+	}
+}
+
+// 7. Fixed-catalog monotonicity: turning any PASS into FAIL never raises any score.
+func TestMonotonicity(t *testing.T) {
+	l := laneMasking()
+	full := allPass(l)
+	top := mustScore(t, rubricOf(l), resOf(full), passing())
+	for id := range full {
+		worse := maps.Clone(full)
+		worse[id] = "FAIL"
+		low := mustScore(t, rubricOf(l), resOf(worse), passing())
+		for _, p := range []string{"global.Q.exact", "lanes.frontend.Q.exact", "lanes.backend.Q.exact"} {
+			if exact(t, low, p).Cmp(exact(t, top, p)) > 0 {
+				t.Fatalf("failing %s raised %s", id, p)
+			}
+		}
+		if exact(t, low, "global.Q.exact").Cmp(exact(t, top, "global.Q.exact")) >= 0 {
+			t.Fatalf("failing %s did not lower global Q", id)
+		}
+	}
+}
+
+// 8. Duplicate control ids and zero weights are invalid input.
+func TestInvalidCatalog(t *testing.T) {
+	zero := base()
+	zero[0] = maps.Clone(zero[0])
+	zero[0]["weight"] = 0
+	for _, tc := range []struct {
+		cs   []map[string]any
+		want string
+	}{{append(base(), base()[0]), "duplicate control id"}, {zero, "must be > 0"}} {
+		if _, err := tryScore(t, rubricOf(tc.cs), resOf(nil), passing()); err == nil || !strings.Contains(err.Error(), tc.want) {
+			t.Fatalf("want %q, got %v", tc.want, err)
+		}
+	}
+}
+
+// 9. Approved sequential degradation is the only open gate -> qualified outcome.
+func TestApprovedDegradation(t *testing.T) {
+	b := base()
+	g := passing()
+	g["gates"].(map[string]any)["G-CONCURRENCY"] = map[string]any{"status": "WAIVED_OPERATIONAL", "approval": "AP-2"}
+	out := mustScore(t, rubricOf(b), resOf(allPass(b)), g)
+	if v := get(out, "readiness_verdict"); v != "MEETS_TARGETS_WITH_APPROVED_DEGRADATION" {
+		t.Fatalf("verdict %s", v)
+	}
+}
+
+// 10. Missing gate evidence -> UNKNOWN -> NOT_READY.
+func TestGatePassWithoutEvidence(t *testing.T) {
+	b := base()
+	g := passing()
+	g["gates"].(map[string]any)["G-COVERAGE"] = map[string]any{"status": "PASS"}
+	out := mustScore(t, rubricOf(b), resOf(allPass(b)), g)
+	if get(out, "gates.G-COVERAGE") != "UNKNOWN" || get(out, "readiness_verdict") != "NOT_READY" {
+		t.Fatalf("gate %s verdict %s", get(out, "gates.G-COVERAGE"), get(out, "readiness_verdict"))
+	}
+}
+
+// 11. A rubric cannot lower the policy thresholds or drop a domain without a recorded exclusion.
+func TestWeakenedRubric(t *testing.T) {
+	b := base()
+	low := rubricOf(b)
+	low["policy"] = map[string]any{"global_min": "0"}
+	var four []map[string]any
+	for i, d := range []string{"tests", "architecture", "ux", "operations"} {
+		four = append(four, ctl(i+1, d, 5))
+	}
+	partial := rubricOf(four)
+	partial["domains"] = map[string]any{"tests": 10, "architecture": 10, "ux": 5, "operations": 5}
+	for _, tc := range []struct {
+		r    map[string]any
+		st   map[string]string
+		want string
+	}{{low, allPass(b), "policy may only raise"}, {partial, allPass(four), "needs a weight or a recorded exclusion"}} {
+		if _, err := tryScore(t, tc.r, resOf(tc.st), passing()); err == nil || !strings.Contains(err.Error(), tc.want) {
+			t.Fatalf("want %q, got %v", tc.want, err)
+		}
+	}
+}
+
+func writeJSON(t *testing.T, dir, name string, v any) string {
+	t.Helper()
+	b, err := json.Marshal(v)
+	if err != nil {
+		t.Fatal(err)
+	}
+	p := filepath.Join(dir, name)
+	if err := os.WriteFile(p, b, 0o644); err != nil {
+		t.Fatal(err)
+	}
+	return p
+}
+
+func runCLI(args ...string) (int, string, string) {
+	var o, e bytes.Buffer
+	code := Run(args, &o, &e)
+	return code, o.String(), e.String()
+}
+
+func TestCompare(t *testing.T) {
+	var excl []any
+	for _, d := range doms[2:] {
+		excl = append(excl, map[string]any{"domain": d, "reason": "out of scope", "evidence": []string{"EV-x"}})
+	}
+	r := map[string]any{"schema": "overhaul.rubric/1", "rev": 1, "lanes": []string{"backend"},
+		"domains": map[string]any{"correctness": 50, "security": 50}, "domain_exclusions": excl,
+		"controls": []map[string]any{ctl(1, "correctness", 5), ctl(2, "correctness", 5), ctl(3, "security", 5)}}
+	dir := t.TempDir()
+	code, stdout, stderr := runCLI("compare", "--rubric", writeJSON(t, dir, "r.json", r),
+		"--baseline", writeJSON(t, dir, "a.json", resOf(map[string]string{"C-001": "PASS", "C-002": "UNKNOWN", "C-003": "PASS"})),
+		"--candidate", writeJSON(t, dir, "b.json", resOf(map[string]string{"C-001": "FAIL", "C-002": "PASS", "C-003": "PASS"})))
+	if code != 0 {
+		t.Fatalf("exit %d: %s", code, stderr)
+	}
+	out := norm(t, json.RawMessage(stdout))
+	if get(out, "baseline_Q.exact") != "15/2" || get(out, "candidate_Q.exact") != "15/2" {
+		t.Fatalf("Q %v / %v", out["baseline_Q"], out["candidate_Q"])
+	}
+	rows := ovio.List(out["transitions"])
+	if len(rows) != 2 ||
+		get(rows[0], "control") != "C-001" || get(rows[0], "kind") != "regression-or-new-failure" ||
+		get(rows[1], "control") != "C-002" || get(rows[1], "kind") != "evidence-only" {
+		t.Fatalf("transitions %v", rows)
+	}
+}
+
+func TestCLIOutWritesScorecard(t *testing.T) {
+	dir := t.TempDir()
+	b := base()
+	outPath := filepath.Join(dir, "scorecard.json")
+	code, stdout, stderr := runCLI("--rubric", writeJSON(t, dir, "r.json", rubricOf(b)),
+		"--results", writeJSON(t, dir, "s.json", resOf(allPass(b))),
+		"--gates", writeJSON(t, dir, "g.json", passing()), "--out", outPath)
+	if code != 0 || stdout != "MEETS_TARGETS -> "+outPath+"\n" {
+		t.Fatalf("exit %d stdout %q stderr %q", code, stdout, stderr)
+	}
+	card, err := ovio.LoadObject(outPath)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if get(card, "schema") != "overhaul.scorecard/1" || get(card, "readiness_verdict") != "MEETS_TARGETS" ||
+		len(get(card, "inputs.rubric_sha256")) != 64 {
+		t.Fatalf("scorecard %v", card)
+	}
+}
+
+func TestCLIInvalidInputExits2(t *testing.T) {
+	dir := t.TempDir()
+	b := base()
+	bad := rubricOf(b)
+	bad["schema"] = "overhaul.rubric/0"
+	rubric := writeJSON(t, dir, "r.json", rubricOf(b))
+	stale := resOf(allPass(b))
+	stale["rubric_digest"] = strings.Repeat("0", 64)
+	for name, args := range map[string][]string{
+		"bad rubric": {"--rubric", writeJSON(t, dir, "bad.json", bad), "--results", writeJSON(t, dir, "s.json", resOf(allPass(b))),
+			"--gates", writeJSON(t, dir, "g.json", passing())},
+		"stale digest": {"--rubric", rubric, "--results", writeJSON(t, dir, "stale.json", stale),
+			"--gates", filepath.Join(dir, "g.json")},
+	} {
+		code, stdout, stderr := runCLI(args...)
+		if code != 2 || stdout != "" || !strings.HasPrefix(stderr, "ERROR: ") {
+			t.Fatalf("%s: exit %d stdout %q stderr %q", name, code, stdout, stderr)
+		}
+	}
+}